The Diagnosis of Dengue in Patients Presenting With Acute Febrile Illness Using Supervised Machine Learning and Impact of Seasonality.

Background: Symptomatic dengue infection can result in a life-threatening shock syndrome and timely diagnosis is essential. Point-of-care tests for non-structural protein 1 and IgM are used widely but performance can be limited. We developed a supervised machine learning model to predict whether patients with acute febrile illnesses had a diagnosis of dengue or other febrile illnesses (OFI). The impact of seasonality on model performance over time was examined. Methods: We analysed data from a prospective observational clinical study in Vietnam. Enrolled patients presented with an acute febrile illness of <72 h duration. A gradient boosting model (XGBoost) was used to predict final diagnosis using age, sex, haematocrit, platelet, white cell, and lymphocyte count collected on enrolment. Data was randomly split 80/20% into a training and hold-out set, respectively, with the latter not used in model development. Cross-validation and hold out set testing was used, with performance over time evaluated through a rolling window approach. Results: We included 8,100 patients recruited between 16th October 2010 and 10th December 2014. In total 2,240 (27.7%) patients were diagnosed with dengue infection. The optimised model from training data had an overall median area under the receiver operator curve (AUROC) of 0.86 (interquartile range 0.84-0.86), specificity of 0.92, sensitivity of 0.56, positive predictive value of 0.73, negative predictive value (NPV) of 0.84, and Brier score of 0.13 in predicting the final diagnosis, with similar performances in hold-out set testing (AUROC of 0.86). Model performances varied significantly over time as a function of seasonality and other factors. Incorporation of a dynamic threshold which continuously learns from recent cases resulted in a more consistent performance throughout the year (NPV >90%). Conclusion: Supervised machine learning models are able to discriminate between dengue and OFI diagnoses in patients presenting with an early undifferentiated febrile illness. These models could be of clinical utility in supporting healthcare decision-making and provide passive surveillance across dengue endemic regions. Effects of seasonality and changing disease prevalence must however be taken into account-this is of significant importance given unpredictable effects of human-induced climate change and the impact on health.

Predicting 3-year mortality based on the tumor necrosis factor alpha concentration in low-flux hemodialysis patients.

Tumor necrosis factor alpha (TNF-α) is an inflammatory cytokine produced during acute inflammation. Few studies have evaluated the association between serum TNF-α and its receptors and their clinical outcomes in hemodialysis patients. However, a study assessing patients using a low-flux dialyzer reuse has not been conducted yet. The serum TNF-α concentrations of 319 prevalent hemodialysis patients (mean age, 45 ± 15 years; median duration of hemodialysis, 48 [interquartile range, 26-79] months; 185 males and 134 females) was examined to predict their 3-year mortality. The patients were divided into tertiles according to their serum TNF-α concentrations: T1 (n = 106; serum TNF-α concentration, <41.22 pg/mL), T2 (n = 106; serum TNF-α level, from 41.22 to 67.28 pg/mL), and T3 (n = 107; serum TNF-α concentration, ≥ 67.29 pg/mL). During the 36-month follow-up period, a total of 50 (15.7%) patients died from all causes. The Kaplan-Meier analysis revealed that the all-cause mortality in T3 was significantly higher compared to that in T1 and T2 (log-rank test, P < .001). The serum TNF-α level was a significant predictor for all-cause mortality (area under the curve = 0.887, P < .001, cutoff value, 89.812 pg/mL, sensitivity = 76%, specificity = 96.3%). The serum TNF-α level was a better predictor of mortality than the duration of hemodialysis and serum albumin, serum high-sensitivity C-reactive protein, and serum beta-2 microglobulin concentrations. The serum TNF-α concentration was a good predictor of the 3-year mortality in low-flux hemodialysis patients.

Clinical course after surgical treatment of active isolated native mitral valve endocarditis.

AIM: Isolated mitral valve endocarditis (MVE) forms a particular subgroup within native infective valve endocarditis (NVE). We characterized this particular subgroup and analyzed the course of patients undergoing cardiac surgery. METHODS: Between 1997 and 2011, 474 patients underwent cardiac surgery at our institution for NVE treatment. Of these, 89 patients (18.8%) suffered from MVE. Valve replacement was undertaken in 84.2% and valve repair in 15.8%. Follow-up was completed with 267 patient years. RESULTS: A delay between the onset of first symptoms and surgery of 4.7±1.2 weeks was observed. Hence, most patients were in a critical preoperative state characterized by severe sepsis and destruction of the mitral valve. About 19.4% were emergency procedures. The MVE group presented with a higher prevalence of preoperative stroke, atrial fibrillation, coronary artery disease and chronic obstructive pulmonary disease in comparison with remaining NVE cases. MVE was more likely caused by Staphylococcus aureus; Staphylococcus epidermidis and Staphylococcus viridans were less frequent (P<0.01 each). Early mortality (6.7%) was caused by persistent sepsis. ICU stay >7 days and time on artificial ventilation >40 h led to a higher risk of in-hospital death. Five-year survival was 59.6% and affected by extracardiac comorbidities. CONCLUSION: Isolated MVE was characterized by a long delay before surgery, differences in microbiological findings and a higher prevalence of preoperative strokes in comparison to NVE. Surgery for MVE can be conducted with good clinical results, but mid-term outcome is limited by extracardiac comorbidities.